Fair Ranking with Noisy Protected Attributes
This addresses the issue of fairness degradation in rankings due to attribute errors, which is crucial for fairness in information retrieval and machine learning applications, though it is incremental as it builds on existing fair-ranking methods.
The paper tackles the fair-ranking problem when protected attributes are noisy, proposing a framework that incorporates fairness constraints and probabilistic perturbation information to maintain utility while ensuring fairness, with provable guarantees and empirical improvements in fairness and fairness-utility trade-off.
The fair-ranking problem, which asks to rank a given set of items to maximize utility subject to group fairness constraints, has received attention in the fairness, information retrieval, and machine learning literature. Recent works, however, observe that errors in socially-salient (including protected) attributes of items can significantly undermine fairness guarantees of existing fair-ranking algorithms and raise the problem of mitigating the effect of such errors. We study the fair-ranking problem under a model where socially-salient attributes of items are randomly and independently perturbed. We present a fair-ranking framework that incorporates group fairness requirements along with probabilistic information about perturbations in socially-salient attributes. We provide provable guarantees on the fairness and utility attainable by our framework and show that it is information-theoretically impossible to significantly beat these guarantees. Our framework works for multiple non-disjoint attributes and a general class of fairness constraints that includes proportional and equal representation. Empirically, we observe that, compared to baselines, our algorithm outputs rankings with higher fairness, and has a similar or better fairness-utility trade-off compared to baselines.